ES2847230T3 - RAMAN spectroscopy system, apparatus and procedure to analyze, characterize and / or diagnose a type or nature of a sample or tissue, such as abnormal growth - Google Patents

Abstract

Raman spectroscopy apparatus (2100), comprising: a first illumination source (2106), configured to direct a collimated illumination output to tissue (2104); a Raman spectrograph (2108), configured to simultaneously detect fingerprint (FP) and high wavenumber (HW) Raman spectra from tissue-scattered illumination; and a computerized control and analysis module (2116), comprising at least one processing unit and memory that stores program instructions executable by the at least one processing unit, to analyze discrete spectral subranges of the Raman spectra in FP and HW wavenumber ranges to identify a match to specific peaks in the detected Raman spectra of one or both wavenumber ranges, characterized by further comprising: an additional illumination source (2118) configured to emit additional illumination to the tissue while the first illumination source directs collimated illumination to the tissue and a heated mirror filter (2120) configured to compensate for illumination interference between the collimated illumination emitted by the first illumination source (2106) and the additional illumination emitted by the additional lighting source (2118).

Description

DESCRIPTION

RAMAN spectroscopy system, apparatus and procedure to analyze, characterize and / or diagnose a type or nature of a sample or tissue, such as abnormal growth

Prior state of the art

Aspects of the present invention relate to a Raman spectroscopy system and a method for improved precision identification of a type or nature of a sample or tissue to which the excitation energy is directed (eg, collimated illumination), such as abnormal or apparently abnormal growth (eg, such as cancer). In particular, but not exclusively, aspects of the present invention allow for improved precision, real-time diagnosis of abnormal tissues such as gastrointestinal growths, in vivo and ex vivo.

It is highly desirable to accurately and rapidly characterize and identify the nature of tissues such as neoplastic gastrointestinal growths by means of minimally invasive techniques, such as endoscopy. For example, colorectal cancer (CRC) is a common disease with a high mortality rate when it is discovered at a late stage. The identification and eradication of neoplastic polyps is one of the most important measures to reduce mortality and colorectal morbidity. Differentiating between hyperplastic polyps with little or no risk of malignant transformation and adenomas with prominent malignancy latency remains a clinical challenge using conventional colonoscopy techniques.

With more than 1.2 million new cancer cases and an estimated 608,700 deaths annually, CRC is a major problem in the modern world. Early identification of precancerous polyps (i.e., adenoma) in curable stages, along with appropriate therapeutic interventions, such as polypectomies or endoscopic mucosal resections (EMR), remain the most important measures to reduce colorectal mortality and morbidity. . Existing colonoscopic approaches suffer from a number of fundamental clinical limitations. This is because conventional colonoscopy relies entirely on visualization of the macroscopic mucosal characteristics of polyps, such as fossa patterns, vascular patterns, etc., and provides little or no biomolecular information about the tissue. Therefore, current standards of care recommend resection of all suspicious polypoid lesions or abnormal growths identified during colorectal examinations. This approach is labor-intensive, leads to expensive histopathological evaluations, and results in unnecessary risk to the patient, as between one-third and one-half of all polypoid lesions turn out to be hyperplastic. Although the absolute risk of polypectomy is considered relatively small, it is still the most common cause of complications, such as bleeding, perforation, etc. during the colonoscopy. Given the existing clinical challenges and the recent introduction of extended population-based colorectal screening programs, the need for advanced endoscopic approaches has never been greater. Therefore, recent research has been directed to the development of more sophisticated molecular spectroscopy and imaging techniques to improve in vivo diagnostics and analysis.

Evidence has confirmed that the application of ex vivo Raman spectroscopy on colon tissue samples provides encouraging accuracies, ranging from 89% to 99%, for discrimination between different disease types (i.e., normal, hyperplastic, adenoma, and adenocarcinoma polyps. ). This still requires removal of the polyps for Raman spectroscopy to identify the nature of the pathological types and therefore has the same problems as the conventional endoscopic techniques mentioned above. Furthermore, there have been many barriers to the translation of Raman spectroscopy in in vivo clinical diagnostics. These include technical limitations such as inherently weak Raman scattering of a tissue, long acquisition times (> 5 s), and a fundamental need to develop long (> 1.9 m) miniaturized fiber optic probe designs with low silica interference. fused, high collection efficiency and depth resolution capability. The probe has to be made with a low content of fused silica, since fused silica has a strong Raman signal and a fluorescence background that will interfere with the weak Raman signal of a tissue. To date, these technical limitations remain unresolved.

Recent technological advances, including the development of fast Raman spectroscopy techniques and miniaturized fiber-optic Raman probes with confocal capability, have enabled real-time in vivo histopathological evaluations (i.e., optical biopsy) during ongoing endoscopy. Raman spectroscopy studies of colorectal polyps have tended to be limited to the so-called fingerprint spectral range (FP) (eg, ^{between 800 and 1800 cm-1). Some attention has been directed towards using a} high-wavenumber (Hw) regime (e.g., between 2800 and 3600 cm-1) as this spectral range presents stronger tissue Raman signals, as well as less interference from the silica background of fiber optic Raman probes. Patent documents WO2014 / 129970 A and US2013231573 A1 disclose examples of such techniques.

At present, however, there are no techniques that allow the identification of cancer cells in vivo that have sufficient precision to make this type of technique a viable option. There is a need for Raman spectroscopic techniques that allow more accurate identification of cancer cells in vivo and ex vivo.

An objective of the embodiments according to the present invention is to overcome at least some of the problems associated with the prior art and current endoscopic techniques to characterize, identify and / or diagnose the type and / or nature of the tissue or tissues such as abnormal growths, for example, cancers and the like in virtually any part of the body. Another objective is to provide an improved precision system and method based on Raman spectroscopy to rapidly characterize, identify and / or diagnose the type or nature of abnormal tissues, such as polyps and precancer, in vivo during endoscopic investigations (e.g. , gastrointestinal endoscopic investigations, such as colonoscopy).

Characteristics of the invention

According to the present invention, a Raman spectroscopy apparatus according to appended claim 1 and a Raman spectroscopy method according to appended claim 8 are disclosed.

Various embodiments according to the present invention describe a system and a method for fiber optic Raman spectroscopy, which provide a labelless vibratory spectroscopic technique that allows a more precise optical biopsy at the biomolecular level in vivo. Multiple embodiments allow a combination of simultaneous FP and HW spectral range measurements during ongoing endoscopy. The rationale for combining FP and Hw spectral ranges for in vivo and ex vivo Raman measurements are diverse:

(i) For tissues that could show strong autofluorescence (e.g. gastric tissue, lung, colon, liver) that exceed tissue Raman signals in the FP range, the hW range may still contain strong tissue Raman peaks with information from diagnosis.

(ii) The FP and HW ranges contain complementary biomolecular information (eg for proteins, lipids, DNA, and water) and thus can enhance tissue characterization and diagnosis.

(iii) The different links vibrate in different spectral ranges, so the use of two different spectral ranges (FP and HW) increases the biomolecular information obtained in a single scan.

According to one embodiment of the present invention, a combined FP and HW fiber optic confocal Raman spectroscopy technique (e.g., involving simultaneous acquisition of FP and HW spectra) can improve real-time diagnosis of cancer, precancer, and / or other abnormal growths in vivo during examinations of the body. The combined FP and HW technique can also be used ex vivo on tissue samples, to more accurately identify the types of abnormal growth present in samples from any part of the body. According to one aspect of the present invention, a Raman spectroscopy apparatus includes: a first illumination source configured to direct illumination toward tissue; a Raman spectrograph, configured to simultaneously detect FP and HW Raman spectra from tissue scattered illumination; and a computerized analysis and control module, comprising at least one processing unit and a memory that stores program instructions executable by the at least one processing unit to analyze discrete spectral sub-ranges (for example, approximately between 3 and 15 or approximately 5 to 10 discrete spectral sub-ranges, where a given one, or each spectral sub-range may have a spectral width of between approximately 10 and 30 cm-1 or approximately 20 cm-1) of the Raman spectra detected in ranges of FP and HW wavelength to identify a match with one or more reference markers in one or both wavelength ranges.

In multiple embodiments, the Raman spectrograph has a single broadband diffraction grating. The first illumination source includes a collimated illumination source to generate excitation energy to apply to the tissue, and the apparatus further includes a probe, to transmit the collimated illumination to the tissue and return the detected Raman spectra from the tissue to the spectrograph. Raman.

The one or more reference markers can include, or be specific peaks in the detected Raman spectra. The computerized analysis and control module may include program instructions executable by the at least one processing unit to diagnose abnormal growth based on the match. The probe may include or be a confocal fiber optic probe. The apparatus may further include an endoscope having an elongated cylindrical body having an instrument channel within which the probe is housed. The computerized control and analysis module may include program instructions executable by the at least one processing unit to dynamically adjust the power of the collimated illumination, and / or dynamically adjust the time during which the tissue is exposed to illumination. collimated.

The apparatus may include a calibration apparatus, configured to standardize the probe or the entire Raman apparatus with respect to at least one calibration reference.

The apparatus may include an additional light source configured to emit additional illumination to the tissue; and a hot mirror filter configured to compensate for illumination interference between the illumination emitted by the first illumination source and the additional illumination emitted by the additional illumination source.

According to one aspect of the present invention, a method performed by a Raman spectroscopy apparatus includes: directing the illumination emitted by a first illumination source to tissue; simultaneously detecting by means of a probe Raman spectra of FP and HW from illumination scattered by the tissue; and analyze discrete spectral sub-ranges in the detected Raman spectra (for example, between approximately 3 and 15 or between approximately 5 and 10 discrete spectral sub-ranges, where a certain one, or each spectral sub-range may have a spectral width between approximately 10 and 30 cm- 1 or approximately 20 cm-1) in the FP and HW wavelength ranges to identify a match with one or more reference markers in one or both wavelength ranges.

Simultaneous detection of FP and HW Raman spectra can include diffraction illumination in the FP and HW wavelength ranges using a single broadband diffraction grating.

The procedure may include diagnosing the nature of an abnormal growth based on the match. The one or more reference markers can include, or be specific peaks in the detected Raman spectra.

The method may further include dynamically adjusting the power of the illumination and / or dynamically adjusting the time during which the fabric is exposed to the illumination.

The procedure may include performing a calibration or standardization procedure to standardize the probe or the entire Raman apparatus against at least one calibration reference prior to illuminating the tissue. The method may further include directing additional illumination toward the tissue using an additional illumination source, while directing the illumination emitted by the first illumination source at the tissue; and compensating for illumination interference between the illumination emitted by the first illumination source and the additional illumination emitted by the additional illumination source using a hot mirror filter.

Reference will now be made, by way of example, to the accompanying drawings to provide a better understanding of embodiments according to the present invention. Drawings should not be construed as limiting and dimensions may not be to scale.

Figure 1 is a schematic diagram of a confocal Raman spectroscopy system for in vivo tissue diagnosis and characterization, according to one embodiment of the invention;

Figure 2 is a block diagram of a wideband fiber optic confocal Raman spectroscopy system, developed for improved precision diagnosis and characterization of tissues during endoscopy, in accordance with one embodiment of the invention;

Figure 3 is a graph of concomitant in vivo autofluorescence and Raman spectra of FP and broadband HW, and obtained from gastrointestinal (GI) mucosa in real time, according to an embodiment of the invention;

Figure 4 is a graph showing an example of interval selection applied to Raman spectra and concomitant broadband in vivo autofluorescence, comprising both fingerprint (FP) and high wave number (HW) areas, according to one embodiment of the invention;

Figure 5 is a graph showing an example of interval selection applied to broadband Raman spectra in vivo after subtraction of the autofluorescence background in the zone (a) between 800 and 1800 cm-1 and (b) between 2800 and 3100 cm-1 respectively, according to an embodiment of the invention;

Figure 6 is a graph showing the Raman spectra plus or minus one standard deviation of three lesions and associated photographs of the three lesions, according to one embodiment of the invention.

FIG. 7A is a difference spectra graph resolving spectral characteristics, according to one embodiment of the invention.

Figure 7B is an analysis of variance (ANOVA) of three categories of tissue over a full spectral range, according to one embodiment of the present invention;

Figure 7C is a histogram of significant Raman peaks, according to one embodiment of the present invention; Figures 8A to D are histopathological sections stained corresponding to different types of colorectal tissue, according to known practices;

Figure 9A is a representation of the posterior probability, according to one embodiment of the present invention.

Figure 9B is a Receiver Operating Characteristics (ROC) chart for diagnosing adenoma and adenocarcinoma, according to one embodiment of the present invention.

Figure 10 is a graph of receiver operating characteristics for distinguishing an adenoma versus benign polyps, according to one embodiment of the present invention;

Figure 11A shows the mean standard deviation of FP / HW Raman spectra in vivo ± 1 of a learning data set (80% of a total data set) for the development of diagnostic algorithms for esophageal squamous cell carcinoma ( CCEE), according to an embodiment of the present invention;

Figure 11B shows the standard deviation of difference spectra (CCEE - normal) ± 1, which resolves the unique spectral characteristics of CCEE, and corresponding images from WLR-guided FP / HW Raman procedures on a normal esophagus and CCEE, according to an embodiment of the present invention;

Figure 12A shows the unpaired two-sided Student's t-test on Raman peak intensities of a learning data set (80% of the total data set) (normal (n = 736); CCEE (n = 202)) on a full spectral range (that is, between 800 and 1800 cm-1 and between 2800 and 3600 cm-1), where multiple (for example, seven) Raman spectra subareas containing the most relevant diagnostic information were identified according to an embodiment of the present invention;

Figure 12B shows a ± 1 SD histogram of the most diagnostic Raman peaks (* p <1E-10), according to an embodiment of the present invention;

Figure 13A shows histopathological sections stained with representative hematoxylin and eosin, corresponding to normal superficial keratinized squamous epithelium and the basal layer;

Figure 13B shows representative histopathological sections stained with hematoxylin and eosin corresponding to invasive esophageal squamous cell carcinoma showing prominent architectural and cytological atypia;

Figures 14A to 14C show posterior probabilities of in vivo Raman spectra belonging to (i) a normal esophagus (n = 736), and (ii) CCEE (n = 202) from the learning data set (80% of the set of total data), using a partial least squares discriminant analysis and a cross-validation leaving a patient out, based on integrated or simultaneous FP, HW and FP / HW Raman techniques, respectively ((O) normal, ( ▲) CCEE) , according to an embodiment of the present invention;

Figure 15 shows receptor operating characteristic (ROC) curves to separate CCEE versus normal esophageal tissue for the learning data set (80% of the total data set), where the areas under the ROC curves (AUC, Areas Under the ROC curves) are 0.972, 0.928, and 0.995, respectively, using integrated or simultaneous Fp, HW, and FP / HW Raman techniques, according to one embodiment of the present invention; Figure 16A is a graph showing composite NIR-AF and Raman spectra measured in cervical patients, according to one embodiment of the present invention.

Figure 16A is a graph showing composite NIR-AF and Raman spectra measured in cervical patients, according to one embodiment of the present invention.

FIG. 16B is a graph showing selected spectral zones extracted after interval PLS-DA, according to one embodiment of the present invention.

Figure 17 is a graph showing a classification error plotted as a function of model complexity for PLS-DA across the spectrum and PLS-DA across ranges, which only uses a ~ 10% fraction of the entire spectrum, according to an embodiment of the present invention;

Figure 18 is a scatter diagram of the posterior probabilities (normal (n = 1001) and precancer (n = 232)) belonging to the precancerous group using PLS-DA modeling (a) across the spectrum and (b) using PLS-DA modeling of intervals, according to one embodiment of the present invention;

Figure 19 is a graph of a continuous Raman spectrum and selected spectral zones after interval PLS-DA measured in 90 gastric patients (benign (n = 1950) and cancer (n = 108)), according to an embodiment of the present invention;

Fig. 20 is a classification error plot represented as a function of model complexity for full spectrum PLS-DA and interval PLS, according to one embodiment of the present invention;

Figure 21 is a scatter diagram of the posterior probabilities (normal (n = 1950) and cancer (n = 108)) of belonging to the cancer group using (a) PLS in the entire spectrum of F ^p and HW y ( b) in PLS- ^d A modeling of intervals, according to an embodiment of the present invention;

Figure 22A shows the mean standard deviation (SD) of Raman spectra of FP and HW in vivo ± 1 of gastric MI (n = 329) and normal mucosa (n = 1083) obtained in 63 patients during a clinical endoscopic examination, according to a carrying out the present invention;

Figure 22B shows difference spectra (ie MI-normal ± 1 standard deviation (SD)) that resolve the unique spectral characteristics between normal and MI gastric tissues, according to one embodiment of the present invention;

Figures 23A and 23B show photomicrographs of hematoxylin and eosin stained sections (H and E) of gastric tissues: A, normal gastric mucosa (X200 magnifications); B, extensive intestinal metaplasia (X100 magnification);

Figure 24 shows the first five principal components (PC, Principal Components) that represent ~ 88% of the total variance calculated from Raman spectra of integrated Fp and HW of gastric tissue (PC1 = 45.6%; PC2 = 33, 6%; PC3 = 4.2%; PC4 = 3.1%; pC5 = 1.2%;), according to an embodiment of the present invention;

Figures 25A, 25B and 25C show scatter diagrams of posterior probability values belonging to MI and normal gastric tissue categories calculated using (A) FP, (B) HW and (C) integrated FP and HW Raman techniques, respectively, using PCA-LDA in conjunction with cross-validation procedures leaving a tissue site out, according to one embodiment of the present invention, where the dotted lines (0.5) provide diagnostic sensitivities of 96.3% (26/27) , 77.8% (21/27) and 92.6% (25/27), and specificities of 87.5% (77/88), 78.4% (69/88) and 90.9% (80 / 88), respectively, using integrated FP, HW, and FP / HW Raman techniques to separate an MI from normal gastric tissue; ((O) normal, (▲) MI),

Figure 26 shows Receiver Operating Characteristics (ROC) curves of classification results to distinguish MI from normal gastric tissue for Raman integrated FP / HW, FP, and HW, respectively, along with PCA-LDA algorithms with validation techniques. crossed leaving a tissue site out, according to an embodiment of the present invention, where the integration areas under the ROC curves are 0.96, 0.94, and 0.79, respectively, for integrated FP / HW Raman techniques, Raman from FP and Raman from HW;

Figure 27 is an example of PLS-DA of intervals applied to Raman spectra measured from the oral cavity, in which an area of great interanatomic variability (for example, 956 cm-1 of hydroxyapatite, 1302 cm-1 and 1445 cm- 1 lipid) is discarded by variable selection techniques, according to one embodiment of the present invention;

Figure 28 is a flow chart of a Raman spectrum processing method for prediction when using discrete spectral ranges, in accordance with one embodiment of the present invention;

Figures 29A and 29B show graphs of an in vivo Raman spectrum obtained in non-contact mode: (a) absence of a hot mirror; (b) integrated with a hot mirror, in front of a xenon light source, according to an embodiment of the present invention;

Figures 30A and 30B show graphs of an in vivo Raman spectrum obtained in contact mode: (a) absence of a hot mirror; (b) integrated with a hot mirror, facing a xenon light source, according to an embodiment of the present invention;

Figure 31 is a block diagram of a generalized illumination light filtering system for any fiber optic Raman spectroscopy application in biomedicine, in accordance with one embodiment of the present invention;

Figure 32 is a flow chart of a calibration procedure, according to one embodiment of the present invention;

Figure 33 is a block diagram of a calibration device used to test and calibrate the fiber optic Raman endoscopy technique prior to use in patients, in accordance with one embodiment of the present invention;

Figure 34 is a diagram of a test routine for Raman endoscopy prior to use in patients, in accordance with one embodiment of the present invention;

Figure 35 is an example graph of background spectra obtained using two different fiber optic probes, in accordance with one embodiment of the present invention.

Figure 36 is a block diagram showing a technique or procedure for calibrating a fluorescence standard glass container in the calibration device using a standard tungsten lamp, in accordance with one embodiment of the present invention;

Fig. 37 is a graph showing an example of calibration functions for two different fiber optic probes, according to an embodiment of the present invention;

Fig. 38 is a graph showing a measurement example of a material with well-defined Raman peaks (eg, polystyrene), according to an embodiment of the present invention;

Fig. 39 is a graph showing a polynomial mapping measurement example of wave number versus number of pixels, according to an embodiment of the present invention;

Figure 40A is a graph showing a comparison of Raman spectra obtained from a two-layer tissue phantom (ie, polystyrene and polyethylene) using 2 different Raman probes, according to one embodiment of the present invention;

Figure 40B is a graph showing a representative Raman spectrum of an upper layer and a lower layer in the tissue phantom, according to an embodiment of the present invention;

Figure 41 is a block diagram of a technique or method for improving the S / N ratio and avoiding CCD saturation by automatically adjusting laser drive power and stacks, in accordance with one embodiment of the present invention;

Figure 42 is a block diagram of a technique or method for improving the S / N ratio and avoiding CCD saturation by automatically adjusting exposure time and accumulations, in accordance with one embodiment of the present invention; and

Fig. 43 is a block diagram showing a representative database structure for storing diagnostic models for different probes and different organs, according to an embodiment of the present invention.

Raman spectroscopy represents a unique optical vibratory technique based on the fundamental principle of inelastic light scattering. When incident laser light induces a polarization change in molecules, a small proportion of incident photons (~ 1 in 108) are inelastically scattered with frequency changes corresponding to the specific active Raman vibratory modes of the molecules in the sample. Different molecules and different bonds vibrate at different frequencies. Raman spectroscopy is, therefore, capable of collecting a large amount of specific biomolecular information from a large number of inter and / or intracellular components, such as proteins, lipids and deoxyribonucleic acids (DNA), water, etc. in tissues.

Figure 1 shows a fiber optic confocal Raman spectroscopy system or apparatus or a fiber optic confocal Raman spectroscope 100, according to an embodiment of the present invention. System 100 is capable of simultaneously performing Raman spectroscopy measurements in the fingerprint spectral range (FP, FingerPrints) and the high wave number regime (HW, High Wave). Obtaining and analyzing combined FP and HW spectral range measurements, according to embodiments of the present invention, allows greater precision in in vivo and ex vivo analysis of biomolecular information, for example, to characterize, identify and / or diagnose tissue or gastrointestinal tissues.

The fiber optic confocal Raman spectroscope 100 includes a collimated illumination source, such as a Near InfraRed (NIR) diode laser 102; a high performance reflective imaging spectrograph 104; a Charge Coupled Device (CCD) camera optimized for NIR 106; and a probe 108 optically coupled to both laser 102 and spectrograph 104 by means of an optical fiber 110. Probe 108 may be carried by an elongated cylindrical body 105 of an endoscope (eg, within a channel of the instrument, provided by the elongated cylindrical body 105). The fiber optic Raman spectroscope 100 may further include a band pass filter 112 for removing background from laser light illumination, and a long pass filter 114 for passing tissue Raman signals while removing scattered laser light. and interference from the bottom of the fiber. A computer or microcontroller system 124 may provide an automated / computerized control and analysis module to monitor aspects of the Raman spectroscope performance and perform Raman spectral analyzes.

In a representative implementation, the near infrared diode laser 102 may have a maximum output of 300 mW and a wavelength of 785 nm, as would be consistent with the device manufactured, for example, by B&W Tek Inc. The NIR laser 102 generates excitation energy at tip 116 of probe 108 which can cause vibration in any species "illuminated" by probe 108 and thereby give rise to a Raman spectrum. Other types of collimated illumination can be used to replace the diode laser 102. Spectrograph 104 can be equipped with thermoelectric cooling, for example, to about -70 ° C.

The spectrograph 104 may be compatible with a device such as the Acton LS785 f / 2, manufactured by Princeton Instruments Inc. The camera 106 may be compatible with a Pixies 400BR eXcelon, manufactured by Princeton Instruments Inc. In such a representative implementation, the spectroscope 100 can obtain in vivo Raman spectra in the spectral range of 400 to 3600 cm-1, with a resolution of approximately 11 cm-1. The devices shown are presented as representative examples and are not intended to be limiting. Atomic emission lines from mercury-argon spectral calibration lamps can be used for wavelength calibration. Lamps may be HG-1 and AR-1 compatible, manufactured by Ocean Optics, Inc., Dunedin, FL. All calibrated wavelength spectra are corrected for system dependence on wavelength, using a tungsten calibration lamp, such as an RS-10, manufactured by EG&G Gamma Scientific, San Diego, CA.

In some embodiments, to measure FP and HW spectra, tissue Raman signals can be measured by successively changing different laser excitation frequencies or by using a dual transmission grating to cover the entire spectral range, (i.e., between -150 and 1950 cm-1; between 1750 and 3600 cm-1) in high resolution, as disclosed in International Patent Application no. PCT / SG2014 / 000063.

Figure 2 shows a further embodiment detailed herein, which is a broadband (eg, 400 to 3600 cm-1) fiber optic confocal Raman spectroscopy system or apparatus 200 that uses a single reflective grating to measure simultaneously spectra of FP and HW in a way that improves the precision of characterization and diagnosis of tissues, for example, in organs and parts of the body accessible endoscopically.

System 200 includes a near-infrared (NIR) diode laser 202 (Aex = 785 nm), a high-performance reflective spectrograph 204 equipped with a thermoelectrically cooled and NIR-optimized charge-coupled device (CCD) camera 206, and a Specially designed 1.8 mm (outer diameter) 208 fiber optic confocal Raman probe. System 200 further includes a computer / microcontroller system 124, configured to provide automated / computerized control and analysis module, to monitor aspects of Raman spectroscope performance and perform Raman spectral analysis. More specifically, the computer / microcontroller system 124 may include one or more processing units, configured to execute memory-resident program instructions to perform specific Raman spectra obtaining and analysis operations, procedures, or processes, according to an embodiment herein. invention. A custom, wide-band, gold-coated reflective grating (for example, 830 g / mm, having a diffraction efficiency> 90% at ~ 800 nm) is incorporated or integrated into the fiber optic confocal 200 Raman system to cover the entire spectral range (i.e. between 400 and 3600 cm-1) with a spectral resolution of ~ 11 cm-1. The fiber optic confocal endoscopic Raman probe 208 is used for both the delivery of laser light and the collection of Raman signals from tissue in vivo. The confocal endoscopic Raman probe 208 has been previously described in International Patent Application No. PCT / SG2014 / 000063, and includes a plurality of 200 µm filter coated beveled collector fibers (NA = 0.22) surrounding a central light supply fiber (200 µm diameter, NA = 0.22). A 1.0 mm miniature sapphire ball lens (NA = 1.78) is attached to the fiber tip of the confocal probe 208 to narrowly focus the excitation light on the tissue, allowing efficient collection of the Raman spectrum of the epithelial lining (tissue depth <200 pm). The fiber optic confocal Raman probe 208 can be inserted into the instrument channel of medical endoscopes and placed in gentle contact with the epithelium for in vivo tissue characterization and diagnosis using a broadband confocal Raman endoscopy technique, based on a embodiment of the present invention.

Figure 3 shows an example of a Raman spectrum and concomitant autofluorescence of FP and broadband HW measured in gastrointestinal mucosa, covering the areas of FP and HW. High resolution tissue Raman peaks are observed on concomitant autofluorescence in the FP range, with tentative molecular assignments, as follows:

• 853 cm-1, which is related to proteins v (C-C),

• 1004 cm-1, which is related to the aromatic ring vs (C-C) of phenylalanine,

• 1078 cm-1, which is related to the v (C-C) of lipids,

• 1265 cm-1, which is related to amide III v (C-N) and or (N-H) of proteins,

• 1302 cm-1, which is related to the twisting and shaking of CH3CH2 proteins,

• 1445 cm-1, which is related to the deformation of 5 (CH2) of proteins and lipids,

• 1655 cm-1, which is related to amide I v (C = O) of proteins, and

• 1745 cm-1, which is related to v (C = O) of lipids.

Strong Raman peaks are also seen in the HW zone, such as:

• between 2850 and 2885 cm-1, which is related to symmetric and asymmetric stretching of lipid CH2s,

• 2940 cm-1, which is related to stretching of protein CH3,

• 3400 cm-1 in the area between 3100 and 3600 cm-1, which is related to the broad Raman band of stretching vibrations of the OH of the water.

This broadband technique according to an embodiment of the present invention allows the FP or HW range or both FP and HW spectral zones to be used simultaneously for tissue analysis, identification, characterization and / or diagnosis, and is therefore especially useful in endoscopically accessible organs showing intense autofluorescence, rapidly saturating the CCD. The computer / microcontroller system 124 can be configured to selectively use the FP spectral zone, the HW spectral zone, or both FP and HW zones together, for tissue analysis, identification, characterization and / or diagnosis, for example, in response to user input directed to a Graphical User Interface (GUI).

One embodiment includes using the HW spectral area for diagnosis if the FP area exceeds the dynamic range of the CCD. In another embodiment, both the FP and HW spectral zones are used for tissue characterization, identification and / or diagnosis. The wideband fiber optic confocal Raman endoscopy platform allows to switch between different spectral zones (eg HW, FP or FP and HW simultaneous) depending on the saturation level of the CCD and / or the type of tissue measured.

A tissue characterization, identification or diagnostic technique or procedure according to one embodiment of the present invention uses the complementary diagnostic information of the broadband FP and HW spectra. In general, the biomedical spectra of tissues are extremely complex. To convert the subtle molecular differences in Raman spectra between different types of tissues into valuable diagnostic information, sophisticated multivariate statistical analysis techniques are required, such as Principal Components Analysis (PCA). This has been widely practiced using full Raman (ie continuous) spectra for diagnosis and characterization of tissues in the FP or HW zones, respectively.

PCA reduces the dimension of Raman spectra by decomposing them into linear combinations of orthogonal components, such as principal components (PC), in such a way that spectral variations in the data set are maximized. Therefore, PCA has been routinely integrated with efficient clustering algorithms such as Support Vector Machine (SVM), Logistic Regression (LR), and Linear Discriminant Analysis (l Da) to the classification of biomedical Raman spectra. PCA is very effective for data reduction and analysis.

Alternatively, Partial Least Squares (PLS) Discriminant Analysis (DA) has been applied to classification problems, encoding class membership of zeros and ones, representing group affinities in an indicator matrix. And appropriate. The PLS-DA employs the fundamental principle of the PCA, but further rotates the components, such as the latent variables (LV, Latent Variables), maximizing the covariance between the spectral variation and the group affinity so that the LVs explain the relevant variations. for diagnostics, rather than the more prominent variations in the spectral data set. In most cases, this ensures that diagnostic-significant spectral variations are retained in the first LVs.

Most multivariate algorithms (eg PCA or PLS-DA) are not originally designed to deal with large amounts of irrelevant spectral variables. In other words, some spectral areas in Raman spectra can have a degrading effect on the diagnostic model, for example, due to large variance, interference, interanatomic variability, etc.

In one embodiment, a novel diagnostic technique or procedure is disclosed, which utilizes the supplemental information of the FP and HW spectral ranges, for example, as obtained by FP and HW spectral measurements performed using a system 200 of Broadband fiber optic confocal Raman spectroscopy. Said embodiment makes use of discrete spectral sub-ranges (for example, of approximately 3 and 15 or of between approximately 5 and 10 discrete spectral sub-ranges, where a certain one, or each spectral sub-interval, can have a spectral width of between approximately 10 and 30 cm- 1 or about 20 cm-1) in the broadband FP and HW Raman spectra, rather than the continuous spectral ranges for diagnostic purposes. Figure 4 shows an example of wideband Raman measurements of the gastro-intestine, in which complementary information has been extracted from the spectral ranges of FP, that is, between 800 and 1800 cm-1, and of HW, that is, between 2800 and 3600 cm-1. Figures 5A-5B show the same Raman spectra after subtraction of 5th order polynomials in the range of FP and HW, respectively, more clearly revealing the specific Raman peaks. In accordance with embodiments of the present invention, it has been discovered that the use of spectral sub-ranges of the FP and HW zone significantly improves or improves the accuracy of endoscopic Raman diagnosis in vivo of abnormal growths, such as precancer and cancer, as will be described later. An embodiment according to the present invention relates to the diagnosis of gastrointestinal abnormal growths, such as colorectal abnormal growths. Raw FP and HW Raman spectra measured from colorectal tissue in vivo represent a combination of weak tissue Raman signals, strong autofluorescence background, and noise. In order to see and analyze Raman signals, the background and noise must be treated or eliminated. Therefore, the raw spectra are pre-processed by a first-order smoothing filter to reduce spectral noise, such as a Savitzky Golay filter, which has a 3-pixel window width. In the FP zone (between 800 and 1800 cm-1), a fifth-order polynomial is found to be optimal for adjusting the autofluorescence background in the noise-smoothed spectrum, and therefore this polynomial is subtracted from the spectrum of Calibrated FP, to produce the tissue Raman spectrum alone. In the HW range (between 2800 and 3600 cm-1) the first order linear fit is found to be optimal to eliminate the weakest autofluorescence reference. Such pre-processing of each received signal is completed in approximately 30 ms, and therefore the processed Raman spectra and diagnostic results can be displayed on a display device, such as a computer screen, in real time.

After pre-processing, the Raman spectra are analyzed for peaks and / or markers. These peaks or markers are then used to estimate, characterize, or determine the nature of the tissue (the nature of injury or abnormal growth) from which the spectra are derived. This can be done by the computer system using an appropriate means of comparison with known control spectra, reference markers, and the like. The term "reference markers", as used herein, is intended to include one or more individual peaks in the Raman spectrum, or indeed the entire Raman spectrum. If one or more Raman reference markers are indicative of a certain nature of abnormal cell growth, a comparison can be made between an obtained spectrum and a reference marker using a lookup table or the like. It will be appreciated that many different types of comparison techniques or procedures can be employed. Once the comparison has been made and the best match has been determined, the type of abnormal cell growth that is present can be indicated to the user of system 200 by any appropriate means. This can include a visual representation on a computer screen and / or an audible message.

In a test of one embodiment of System 200, a total of 50 consecutive symptomatic patients were colonoscopically examined. The patients had presented for examinations, for surveillance, or screening, for various colorectal indications, such as anemia, bleeding, etc. Before colonoscopy, patients were given a polyethylene glycol electrolyte (PEG) bowel prep. Sedation was performed using propofol administered intravenously. The endoscopists cleaned the colon during inspection and prior to confocal Raman scans, then flat, polypoid colorectal lesions were flushed with physiological saline to further reduce confounders (i.e. stool and residual fluids, etc. ). During a routine examination, the endoscope was directed towards the distal colon, and a Raman scan (ñ 15 spectra) was performed on suspicious lesions during removal of the probe from the body. Each Raman measurement of the tissue was obtained in a period of between approximately 0.1 and 0.5 seconds. This allowed a quick study of colorectal polyps. Any Raman spectra obtained without contact with colonic polyps (~ 10%) were automatically discarded using online clinical software using Principal Component Analysis (PCA) procedures associated with Hotelling's Q and T (i) 2 residue statistics. PCA procedures are subject to International Patent Application No. PCT / SG2014 / 000063 cited above, and they work as follows:

• A new principal component (PCA) -based outlier detection scheme is introduced, coupled with Hotelling's Q and T2 residual statistics, to serve as a high-level model-specific feedback tool in the online framework. . Hotelling's Q and T2 residuals are the two independent parameters that provide information inside and outside the model fit.

• Using Hotelling's Q and T2 residue parameters as indicators to control the quality of the acquired spectrum (ie, probe-tissue contact mode, probe manipulation variations, white light interference, blue light interference, factors of confusion, etc.), the audit feedback has been integrated into the online Raman diagnostic system, facilitating real-time spectroscopic scanning and probe handling advice for clinicians.

• If the spectra are verified for further analysis, they are fed into probabilistic models for in vivo cancer diagnoses. The software can instantly toggle between different pre-rendered multivariate statistical models, including partial least squares discriminant analysis (PLS-DA), linear discriminant analysis (LDA) of pCa, ant colony optimization LDA (Ac O) -LDA (Ant Colony Optimization), Classification and Regression Trees (CART), Support Vector Machine (SVM), Adaptive Boost (Ada-Boost) etc., based on spectral databases of a large number of patients.

After the Raman scan was completed and the results saved, each tissue sample was removed; fixed in formalin; was sectioned; stained with hematoxylin and eosin (H and E); and was sent for histopathological examination. Colorectal tissues were classified into the following three clinically important categories:

(i) benign (normal and hyperplastic polyps)

(ii) adenomas (tubular, tubulovillous, villous) of low and high grade and,

(iii) adenocarcinomas.

The simultaneous FP and HW fiber optic confocal Raman technique, according to one embodiment of the present invention, was compared with histopathological evaluations to determine the ability to differentiate neoplastic from non-neoplastic colorectal lesions, in vivo.

The comparison included the use of statistical analysis of the results to validate whether the simultaneous FP and HW fiber optic confocal Raman spectroscopy technique was sufficiently accurate to replace the currently used techniques. Cohen's k statistics were calculated to assess agreement for histopathological characterization. Analysis of variance (ANOVA) with Fisher's post hoc Least Significant Differences (LSD) test was used to test for differences in means between groups. Multivariate statistical analysis was used to extract significant Raman spectral characteristics for clinical diagnoses. A probabilistic partial least squares (PLS) discriminant analysis (DA) was applied for tissue diagnosis. Cross-validation "leaving a patient out" was used to evaluate and optimize the complexity of the PLS-DA model, reducing the risk of overfitting. Receiver operating characteristic curves (ROC) were generated and the area under the curves (AUC) was calculated to assess the ability of the FP and HW fiber optic confocal Raman spectroscopy technique to differentiate neoplastic from non-neoplastic polyps, in alive.

In an experiment, presented by way of example only, to show the results of using an embodiment, according to the present invention, in a specific group of test subjects or patients, fifty patients (27 men and 23 women) with a mean age / range of 52 / (between 23 and 83) were enrolled for the confocal fiber optic Raman examination. Thirteen patients had low-grade dysplasia with hidden adenomas (eleven tubular adenomas and two tubulovillous). Three patients were associated with advanced-stage colorectal adenocarcinoma. A Cohen's kappa of 0.89 demonstrated a high level of agreement between the pathological findings for the three groups of tissues. A total of 1731 colorectal Raman spectra were successfully obtained in vivo from 126 abnormal growths or lesions. Of these lesions, 1,397 were benign, 235 were adenomas, and 99 were adenocarcinomas. This was confirmed by histopathological examinations.

The detailed distribution of patients and lesions, including pathological subtypes and anatomical locations, such as ascending, transverse, descending, sigmoid, rectum, are summarized in Table 1.

Figure 6 shows the mean standard deviation (SD) of in vivo confocal Raman spectra ± 1 (shaded in light gray), measured for benign, adenoma, and adenocarcinoma abnormal growths. Prominent tissue Raman peaks can be seen, with provisional assignments in the FP range of approximately:

• 853 cm-1, which is related to proteins v (C-C),

• 1004 cm-1, which is related to the aromatic ring v ^s (CC) of phenylalanine,

• 1078 cm-1, which is related to v (C-C) of lipids,

• 1265 cm-1, which is related to amide III v (C-N) and S (N-H) of proteins,

• 1302 cm-1, which is related to the twisting and agitation of CH2 of lipids

• 1445 cm-1, which is related to the deformation of 5 (CH2) of proteins and lipids,

• 1618 cm-1, which is related to v (C = C) of porphyrins

• 1655 cm-1, which is related to the amide I v (C = O) of proteins, and

• 1745 cm-1, which is related to the v (C = O) of lipids.

Likewise, intense Raman peaks are observed in the HW zone such as:

• between 2850 and 2885 cm-1, which is related to the symmetric and asymmetric stretching of lipid CH2, respectively,

• 2940 cm-1, which is related to the stretching of CH3 in proteins,

• 3009 cm-1, which is related to the asymmetric stretching of the = CH of proteins,

• ~ 3300 cm-1, which is related to Amide A (NH stretching of proteins), as well as

• ~ 3200 and ~ 3400 cm-1 in the broad Raman band of water, which is related to stretching vibrations of OHs that are directly related to local conformation and interactions of OH bonds in the intracellular and extracellular space of the tissue.

Figure 7A shows the difference spectra for benign adenoma, benign adenocarcinoma, and adenomaadenocarcinoma ± 1 SD (light gray shading) that resolves distinctive spectral features such as peak intensity, band shift and widening, etc. between the different colorectal lesions.

Figure 7B shows a logarithmic graph of the calculated p-values (ANOVA) for each of the Raman intensities in the entire spectral range (each Raman spectrum varying between 800 and 1800 cm-1 and between 2800 and 3600 cm-1 with a set of 779 intensities).

Figure 7C shows a histogram of the most significant Raman peak intensities for the diagnosis (mean ± 1 SD) of the FP and HW ranges around (i) 1078 cm-1 (ii) 1265 cm-1 (iii) 1335 cm-1 (iv) 1431 cm-1 (v) 1618 cm-1 (vi) 2885 cm-1 and (vii) 3200 cm-1.

The in vivo Raman spectra were correlated with representative histopathological sections of the different pathology categories. Figure 8A shows the extensive presence of goblet cells in normal intestinal-type mucosa with lympho-plasmocytic infiltrate. Figure 8B shows polypoid features with star-shaped glands in a hyperplastic goblet-like polyp. Figure 8C shows a tubular adenoma with low-grade dysplasia showing crowded cells with hyperchromatic nuclei. Figure 8D shows an invasive adenocarcinoma with prominent cellular and architectural abnormalities. Histopathological characterizations in Figures 8A-D reveal the cellular and morphological characteristics of different types of lesions, while the FP and HW fiber-optic confocal Raman spectroscopy technique reveals the biomolecular changes that occur in the epithelium, associated with carcinogenesis. colorectal shown in Figures 7A-C.

Using the complementary biomolecular information from the FP and HW spectral ranges, fiber optic Raman spectroscopy is applied for in vivo diagnosis. Figure 9A shows the a posteriori probabilities generated using the trichotomic probabilistic PLS-DA of the 50 colorectal patients belonging to (i) benign (n = 1397), (ii) adenoma (n = 235) and (iii) adenocarcinoma (n = 99), respectively, using the FP and HW fiber optic confocal Raman spectroscopy technique, according to an embodiment of the present invention. The relationships between sensitivities and specificities were determined by developing the ROC curves, as shown in Figure 9B. The areas under the curves (AUC) are 0.930 and 0.978, to distinguish adenoma from benign polyps, and adenocarcinoma from benign adenoma and polyps, respectively. ROC analysis shows that adenoma can be distinguished from benign flat and polypoid lesions with a sensitivity of 88.5% (208/235) and a specificity of 80.0% (1118/1397). Adenocarcinomas were detected with a sensitivity of 92.9% (92/99) and a specificity of 96.51% (1575/1632) for each spectrum. An additional ROC analysis, as shown in Figure 10, was performed to examine whether the biomolecular information from the FP and HW regimens was complementary to the diagnosis. The AUC was 0.908, based on the FP range (that is, between 800 and 1800 cm-1) and 0.895 based on the HW range (that is, between 2800 and 3600 cm-1). On the other hand, when using the complementary information from the FP and HW ranges, the robustness was significantly improved, leading to an AUC of 0.930. In light of these encouraging in vivo results, the unique FP and HW fiber optic confocal Raman spectroscopy technique represents a very powerful optical modality, allowing real-time objective diagnosis of colorectal neoplasia in vivo.

A large group of people are known to have hidden small colorectal hyperplastic polyps or flat polypoid lesions, which significantly increase the cost of histopathological evaluations. The most important motivation to develop and adopt advanced endoscopic modalities for colorectal examinations is the ability to differentiate benign hyperplastic polyps from an adenoma. The embodiments according to the present invention demonstrate that, for the first time, high quality in vivo confocal Raman spectra covering the FP and HW spectral ranges can be measured (eg, simultaneously) from colorectal polyps, and analyzed in real time, and this can be used to offer an improved way of identifying neoplasms such as those shown in Figure 6. The confocal fiber optic Raman spectroscopy technique according to embodiments of the present invention uncovers a plurality of biochemical changes and biomoleculars occurring in the epithelium, accompanying neoplastic transformation, as shown in Figures 7A-C. For example, neoplastic polyps were associated with significantly lower Raman peak intensities by 1078 cm ^-1 with respect to v (CC); 1431 cm ^-1 with respect to 5 (CH ² ); and 2850 and 2885 cm ^-1 with respect to the symmetric and asymmetric stretching of the CH ² bond, respectively (p <0.001), which points to a relative reduction in lipid content. The results also show a significant upregulated protein, as indicated by the sensitive biomarkers in 1004 cm ^-1 with respect to the aromatic ring v ^s (CC) of phenylalanine, and the widening of the band in 1655 cm ^-1 with with respect to the amide I v (C = O) of the proteins, which points to a greater cell proliferation in the neoplastic epithelium. The intensity of the prominent Raman peak in neoplastic polyps at 1618 cm ^-1 (p <0.001) is also closely related to the hemoglobin content associated with antigenic onset and the resulting neovasculation.

The above-mentioned peak intensities are related to vibrations of specific bonds within molecules. Each type of bond has a different peak intensity, and identifying a peak intensity that can be associated with a specific bond can reveal biomolecular information. The fact that a lesion or abnormal growth includes a specific link (s) and gives rise to a specific associated intensity peak can be used to differentiate between different types of abnormal growth. For example, if an abnormal growth that is benign shows a specific intensity peak, later detection of that intensity peak can lead to the determination that in the subsequent situation the injury, or abnormal growth, is also likely to be benign. .

According to embodiments of the present invention, the link between peak intensity and biomarkers can be exploited to identify the type of abnormal growth that is present. This can occur both in vivo and ex vivo.

In some results, the water content was found to be markedly different in the adenoma. The increase in the intensity of the broad antisymmetric vibration of the OH bond stretch by 3200 cm ^-1 (p <0.001) indicates that the neoplastic epithelium presents an increase in interstitial water content that could be partially explained by the expression of aquaporins that alter permeability to water, thereby inducing hydration of neoplastic cells. The increase in interstitial water may be interrelated with the simultaneous decrease in hydrophobic lipids. The significance of this finding was analyzed by calculating the peak intensity ratio (ie, I ²⁸⁸⁵ / I ³²⁰⁰ ) associated with lipids and interstitial water. This peak ratio alone can distinguish adenoma from benign polyps with a sensitivity of 81.3% (191/235) and a specificity of 80.4% (1132/1397). Therefore, lipid content and water perfusion in polyps are very useful biomarkers for colorectal neoplasia in situ. The direct correlation of epithelial Raman spectral fingerprints with cellular and tissue biochemistry can therefore deepen the understanding of colorectal carcinogenesis at the biomolecular in situ level. The use of spikes or markers can be used to diagnose the type of abnormal growth that is present.

Taking advantage of the wide range of complementary optical biomarkers including proteins, lipids, DNA, and the non-interstitial and interstitial water conformations in a protein, it has been shown for the first time that an accurate diagnosis of adenoma can be made in vivo, as shown in Figures 9A-B. Adenoma could be distinguished from benign flat and polypoid lesions with a sensitivity of 88.5% (208/235) and a specificity of 80.0% (1118/1397), for each spectrum. Although the fiberoptic confocal-type Raman probe selectively interrogates the shallow epithelial layer, it was still able to efficiently discriminate adenocarcinomas from adenomas, meaning that invasive cancer cells are associated with biomolecular abnormalities.

The efficacy of the FP and HW fiber optic confocal Raman colonoscopy technique for in vivo detection and diagnosis of colorectal neoplasia is clear from the results. It has also been shown that using the FP and HW spectral ranges yields an AUC that was superior to using either FP or HW ranges individually. These substantial results establish that the FP and HW Raman spectroscopy technique can greatly reduce the rate of misclassification, confirming the addition of complementary biomolecular information from the FP and HW ranges to improve colorectal diagnosis in vivo. For example, the spectral range of HW contains information related to the local conformation of water, as well as the stretching fractions of the CH ² and CH ³ bond that are not contained in the range of FP. The combination of FP and HW ranges can also be used for ex vivo diagnostic purposes.

It should be noted that the combination of FP and HW Raman spectroscopy techniques is usually a preferred operating procedure, such as by means of simultaneous FP and HW spectral measurements. However, the use of FP or HW in vivo produces results that are capable of identifying different types of abnormal growth. As a result, embodiments according to the present invention can use FP alone, HW alone, or a combination of FP and HW (eg, simultaneously).

Also, it should be noted that the use of discrete spectral sub-ranges (which may also be referred to as default values and / or reference markers) in broadband FP and HW spectra provides an improved diagnostic procedure or technique.

The use of a single gold-coated broadband reflective grating (e.g. to simultaneously obtain FP and HW spectra) is an important development in equipment for higher precision objective characterization and diagnosis of tissue, e.g. to characterize / diagnose potentially abnormal, or abnormal growths in vivo, with greater precision in real time during an endoscopic procedure. The unique gold-coated broadband reflective grating means that there is no need to switch between diffraction gratings during in vivo measurements, as has been the case in the prior art. Obviously, this has many advantages. Therefore, a system or device according to an embodiment of the present invention is capable of performing all necessary measurements without requiring change processes. The device is more compact and cost-effective as it has only one rack instead of multiple racks. Although the single grating design is somewhat compromised in spectral resolution compared to a dual grating design, this compact system would not or would not significantly affect diagnostic purposes or results as the spectral resolution of the compact system matches the Raman spectral bandwidth of the tissue, which is typically in the range of ~ 10 cm-1.

Fiber optic confocal Raman spectroscopy provides real-time, uninterrupted computerized objective diagnosis that is simple to operate and does not require additional endoscopic training or administration of contrast agents. By allowing functional and biomolecular / biochemical evaluation of intestinal epithelium in vivo, the introduction of FP and HW fiber optic confocal Raman spectroscopy will have a major impact on the practice of gastrointestinal endoscopy, such as colonoscopy. This new biomolecular endoscopic approach enables immediate and objective decision making during colorectal clinical examinations. There may be two key functions for fiber optic confocal Raman spectroscopy in colorectal examinations, as follows:

(i) Preventive and interventional approaches, including identification of small high-risk adenomas for immediate or REM polypectomy. Hyperplastic polyps or suspicious flat lesions that are clearly low-risk in nature can be left in situ, efficiently reducing medical costs;

(ii) Fiber optic confocal Raman spectroscopy can also be used to efficiently confirm or reject the presence of colorectal adenocarcinoma with a high degree of precision.

The embodiments according to the present invention open the possibility of their use in endoscopic and laparoscopic surgery of RCC, therefore offering the gastroenterologist an objective tool for the evaluation and definition in real time of the resection margins, as well as the evaluation of the monitoring of post-treatment efficacy or recurrence at the molecular level. This can aid in complete excision of the tumor and subsequent evaluation of the margins to reduce the risk of recurrence. Thus, FP and HW fiber optic confocal Raman spectroscopy according to embodiments of the present invention can be used in the field of gastrointestinal endoscopy, as well as other body sites, both in detection settings and in therapeutic colorectal applications. , by allowing objective evaluations of the tissue in vivo in real time.

Another clinical example used an embodiment, according to the present invention, for real-time in vivo diagnosis of esophageal squamous cell carcinoma (SCC) during endoscopy by simultaneously obtaining both fingerprint Raman spectra (FP) (i.e., between 800 and 1800 cm-1) and high wave number (HW) (that is, between 2800 and 3600 cm-1) of esophageal tissue in vivo. In this set of experiments, a total of 1,172 high-quality in vivo FP / HW tissue Raman spectra (normal (n = 860); CCEE (n = 312)) were obtained from 48 esophageal patients undergoing routine endoscopic examination. The total in vivo Raman data set was divided into two parts: that is, 80% of the total data set was for learning (938 FP / HW Raman spectra in vivo [normal (n = 736); CCEE (n = 202)] of 34 esophageal patients); while the remaining 20% of the total data set was for predictive tests (234 Raman spectra of fP / HW in vivo [normal (n = 124); CCEE (n = 110)] of 14 esophageal patients).

Figure 11A shows the mean standard deviation (SD) of in vivo FP / HW tissue Raman spectra ± 1 (shaded area) of the learning data set (80% of the total data set) for the development of diagnostic algorithms. of tissues. Corresponding images of WLR-guided FP / HW Raman procedures are also shown. Prominent Raman peaks of esophageal tissue can be seen with provisional assignments in the FP zone, i.e .:

• 853 cm-1, which is related to proteins v (C-C),

• 1004 cm-1, which is related to the aromatic ring of phenylalanine,

• 1078 cm-1, which is related to the v (CC) of lipids,

• 1265 cm-1, which is related to amide III v (C-N) and S (N-H) of proteins,

• 1302 cm-1, which is related to the twisting and agitation of lipid CH2,

• 1335 cm-1, which is related to the CH3CH2 twist of proteins and nucleic acids,

• 1445 cm-1, which is related to the deformation of the 5 (CH2) of proteins and lipids,

• 1618 cm-1, which is related to the v (C = C) of the porphyrins

• 1655 cm-1, which is related to the amide I v (C = O) of proteins, and

• 1745 cm-1, which is related to the v (C = O) of phospholipids.

Also, strong Raman peaks are observed in the HW zone as follows:

• between 2580 and 2885 cm-1, which is related to the symmetric and asymmetric stretching of lipid CH2, • 2940 cm-1, which is related to the stretching of protein CH3,

• ~ 3300 cm-1, which is related to amide A (NH stretching of proteins), and

• the broad Raman band of water (stretching vibrations of OHs that reach a maximum between ~ 3250 and ~ 3400 cm-1), which are related to the local conformation and interactions of OH bonds in the intracellular and extracellular space of the esophageal tissue.

Figure 11B shows the difference in Raman spectra between CCEE and normal esophageal tissue ± 1 SD (shaded area), reflecting the changes in active Raman component associated with cancer progression in the esophagus. The significant difference (p = 1.3E-8, unpaired two-sided Student's t-test) in the Raman spectra of a CCEE and discerned normal tissue demonstrates the potential of simultaneous FP / HW Raman endoscopy for the in vivo diagnosis of esophagus cancer.

To elucidate the active Raman components of diagnostic significance, Figure 12A shows a logarithmic plot of the calculated p-values (unpaired two-sided Student's t-test) for each of the Raman intensities over the entire spectral range (i.e., between 800 and 1800 cm-1 and between 2800 and 3600 cm-1). Specifically, the following spectral subzones were found with a statistically significant difference (p <1E-10) between a CCEE and a normal esophagus: between 840- and 940 cm-1, between 1025 and 1100 cm-1, between 1310 and 1355 cm -1, between 1585 and 1690 cm-1 and between 2830 and 2975 cm-1 related to proteins, lipids and nucleic acids. Significant spectral differences were also observed in pore water in the ranges between 3160 and 3260 cm-1 and between 3370 and 3420 cm-1.

Figure 12B shows a histogram of the most statistically different Raman peak intensities (mean ± 1 SD) for the FP and HW ranges, that is, (i) 853 cm-1, (ii) 1078 cm-1, (iii ) 1335 cm-1, (iv) 1618 cm-1, (v) 1655 cm-1, (vi) 2850 cm-1, (vii) 2885 cm-1, (viii) 3250 cm-1 and (ix) 3400 cm-1. As indicated in Figure 13, histopathology identifies prominent cellular and architectural abnormalities in a CCEE, relatively higher or lower FP / HW tissue Raman bands representing different active Raman components reveal specific biochemical / biomolecular changes of esophageal tissue accompanied by transformation of CCEE. Changes in FP / HW Raman spectra related to lipids, proteins, DNA, and water content in tissue reconfirm the ability of simultaneous FP / HW Raman spectroscopy to detect CCEE at the molecular level.

Taking advantage of the complementary biochemical / biomolecular information identified in the spectral ranges of FP and HW, PLS-dA and LOPCV were implemented in the learning data set (80% of the total data set), to develop a robust diagnostic model to improve the diagnosis of CCEE in vivo. A Cohen's kappa of 0.91 demonstrated a high level of agreement among independent pathologists for the esophageal tissue clusters. Figure 14 shows the cross-validated PLS-DA posterior probability scatterplots of each Raman prediction for integrated (a) FP, (b) HW, and (c) FP / HW, respectively. The diagnostic precision with integrated FP / HW Raman spectroscopy is 97.3% [sensitivity 97.0% (196/202) and specificity 97.4% (717/736)], superior to the use of the Raman technique. only of FP (precision-90.9%; sensitivity-93.6% (189/202), and specificity-90.2% (664/736)) or only of HW (precision-85.5%; sensitivity- 78.2% (158/202) and specificity-87.5% (644/736)).

The receiver operating characteristics (ROC) curves were also generated as shown in figure 15, with the integration areas under the rOc curves of 0.972, 0.928 and 0.995, respectively, for the FP, HW and FP / techniques. Integrated HW. The above results confirm that the integrated or simultaneous FP / HW Raman technique provides the best diagnostic results for the detection of CCEE in vivo compared to the FP-only or HW-only Raman technique.

In light of these promising diagnostic results, the developed simultaneous FP / HW Raman spectroscopy and diagnostic algorithms were applied to the predictive diagnosis of the independent test data set (20% of the total data set). Predictive accuracy of 93.2% [ie, sensitivity of 92.7% (102/110) and specificity of 93.6% (116/124)] can be achieved with integrated FP / HW Raman spectroscopy, which demonstrates the advantages of integrated FP / HW Raman spectroscopy over Raman technique FP alone (predictive accuracy-91.0%; sensitivity-90.9% (100/110) and specificity-91.9% (113/124)) or HW alone (predictive accuracy-80.3%; sensitivity -76.4% (84/110), and specificity -83.9% (104/124)) for CCEE in vivo. In another embodiment, the identification of other types of cancer was considered, for example, using the system 200 of Figure 2, to identify cervical cancers or other types of abnormal growths.

To select the complementary spectral ranges of the FP and HW spectral regions, variable / feature selection techniques are incorporated into the broadband Raman endoscopy technique. The benefits of feature / variable selection include:

1. improve predictive performance;

2. reduce the complexity of the model; and

3. gain knowledge about the underlying spectroscopic process, such as the importance of variables / characteristics.

In one embodiment, an interval PLS-DA algorithm is used, but in principle other, multiple, or all of the feature / variable selection techniques or procedures can be applied to select the complementary spectral zones for any clustering algorithm, such as PCA-LDA, SVM, LR, etc. Trait / variable selection techniques could be genetic algorithms (GA), swarm intelligence, selectivity relationships, etc. Briefly, the interval PLS-DA used in this case performs a sequential and exhaustive search for the best combination of intervals in the Raman spectra. Therefore, the interval PLS-DA creates individual PLS models, each of which uses only a subset or window of variables. If there are 200 intervals for a given spectral data set, 200 PLS-DA models are calculated (that is, one for each interval).

Cross-validation is performed leaving one patient out for each model and the interval that provides the highest diagnostic precision is selected. This is the optimal single interval model. If only one interval is desired, the algorithm stops at this point. However, if more than one interval is desired (to increase information content and improve predictive performance), additional cycles can be performed. In the second cycle, the first selected interval is used in all models, but is combined with each of the other remaining intervals, one at a time, creating a new set of new PLS models. In this way, the areas of complementary diagnostic value of the FP and HW range are extracted, while the redundant or irrelevant spectral ranges, such as areas of low predictive power, are excluded from the model. In certain applications, other spectral ranges (for example, the so-called Raman quiet range, between 2000 cm ^-1 and 2800 cm ^-1 ) other than the FP and HW ranges can be used.

To demonstrate an application of this embodiment, a broadband spectrum was obtained from a series of cervical patients. Forty-four non-pregnant female patients (between 18 and 70 years of age) who underwent a colposcopy procedure due to an abnormal Pap smear were recruited. Prior to in vivo tissue Raman spectral measurements, a 5% acetic acid solution was applied topically to the cervix for 2 minutes to assess tissue color whitening (the degree of white discoloration in the cervix is related to the degree of precancer).

Concomitant autofluorescence and confocal broadband Raman spectra were measured by placing the fiber optic confocal probe in light contact with the tissue. Figure 16A shows the raw composite average Raman and autofluorescence spectra of normal (n = 356) and precancerous (n = 120) cervical tissues. Spectra of normal and precancerous cervical tissue show faint vibrating Raman bands at the top of the autofluorescence, near:

• ~ 854 cm ^-1 , which is related to the aromatic deformation of glycogen (CCH) and the stretching of the (CC) of structural protein and collagen,

~ 937 cm ^-1 , which is related to stretching of the v (CC) of proline, valine, and glycogen,

~ 1001 cm ^-1 , which is related to the aromatic ring (CC) of phenylalanine,

~ 1095 cm ^-1 , which is related to phospholipids and nucleic acids),

~ 1253 cm ^-1 , which is related to amide III,

~ 1313 cm ^-1 , which is related to the torsion mode of CH ³ CH ² of lipids / proteins (collagen), ~ 1445 cm ^-1 , which is related to the bending mode of CH ² of proteins and lipids, ~ 1654 cm ^-1 , which is related to the stretch mode of the (C = O) of the amide I band of proteins,

~ 2946 cm ^-1 , which is related to ^{protein CH 3} stretching, and

~ 3400 cm ^-1 , which is related to the stretching of the (OH) of the water.

To demonstrate that the integrated broadband Raman technique with interval selection can improve the diagnosis of cervical precancer, a conventional PLS-DA is applied to the continuous spectrum (Figure 16A) and interval PLS-DA is applied to extract areas. complementary as described above. Fig. 16B shows the selected spectral subzones (that is, between 1000 and 1020 cm-1, between 1640 and 1660 cm-1, between 2890 and 2910 cm-1, and between 3290 and 3310 cm-1) after the interval PLS-DA.

Figure 17 shows the classification error as a function of the complexity of the algorithm (that is, the number of LVs) for both the PLS-DA and PLS-DA models of intervals. The minima can be found at 9 and 5 ^l V respectively. It is evident that by using those complementary spectral zones of the FP and H ^w zone instead of the whole spectrum, the precision increases significantly (the classification error is reduced from 25% to 15%). This example shows that by removing the areas containing little information and selecting the complementary sub-ranges of the FP and ^h W spectra, the precision for in vivo detection of precancer can be significantly improved by approximately 10%. In fact, this proves that the FP and HW zones contain complementary information for tissue characterization. The scatter plots in Figures 18A and 18B show the results of the probabilistic classification. Consequently, the discrete ranges of the broadband confocal Raman endoscope technique now allow clinicians to obtain a more accurate probabilistic measure of the risk associated with suspicious lesions, thereby significantly improving the guidance of physical biopsies.

It should be noted that thresholds can be imposed on probabilistic classifications (Figures 18A and 18B) by incorporating prior information, such as family history of disease. For example, if a patient belongs to a high-risk group, the device can automatically select a threshold that provides greater sensitivity for a precancer or cancer. Other types of data to generate prior probabilities could be genetic, proteomic, epidemiological, such as age, ethnicity, gender, drinking habits, etc., the result of imaging (e.g. CT, MRI, etc.), symptoms, etc. The option to adjust the thresholds has been integrated into the clinical Raman endoscope software, which controls the spectral measurements.

In a further clinical example using the broadband confocal Raman endoscope, fiber optic confocal Raman spectra were obtained from a series of gastric patients, focusing on the early diagnosis of gastric malignancies. Figure 19A shows the mean in vivo confocal FP and HW Raman spectra measured in 90 patients presenting different tissue types: benign (n = 1950 spectra) and cancer (n = 108 spectra) as confirmed by histopathological characterization. Prominent tissue Raman peaks of proteins, DNA, and lipids can be seen around:

936 cm-1, which is related to proteins v (C-C),

1004 cm-1, which is related to the aromatic ring v ^s (CC) of phenylalanine,

1078 cm-1, which is related to the v (C-C) of lipids,

1265 cm-1, which is related to amide III v (C-N) and S (N-H) of proteins,

1302 cm-1, which is related to the twisting and shaking of the CH2 bond of proteins,

1445 cm-1, which is related to the deformation of the 5 (CH2) of proteins and lipids,

1618 cm-1, which is related to the v (C = C) of the porphyrins

1655 cm-1, which is related to the amide I v (C = O) of proteins,

1745 cm-1, which is related to the v (C = O) of lipids.

~ 2946 cm-1, which is related to protein CH3 stretching, and

~ 3400 cm-1, which is related to the stretching of the (OH) of the water.

To demonstrate an embodiment according to the present invention, conventional PLS-DA is applied to the continuous spectrum, as well as interval PLS-DA. Figure 19B shows the complementary spectral zones selected after interval PLS-DA (i.e., between ~ 1050 and 1120 cm-1, between ~ 1323 and 1490 cm-1, and between ~ 2850 and 2870 cm-1).

Figure 20 shows the classification error as a function of algorithm complexity (ie LV number) for both PLS-DA and PLS-DA of intervals. Minimal errors in classification can be found using 5 LVs for both models. It is evident that by choosing those significant spectral zones of complementary diagnosis within the range of FP and HW instead of the entire spectrum, the model becomes more accurate (the classification error is reduced from 0.25% to 0.18% ), as shown in the scatter plot in Figures 21A and 21B. Sensitivity increases from 72.2% to 75.9% and specificity increases from 74.7% to 87.9%, showing that supplementary information improves the diagnosis of gastric cancer.

Yet another clinical example used an embodiment according to the present invention in which a simultaneous fiber optic Raman endoscopic technique of FP and HW was used for the in vivo detection of precancerous gastric intestinal metaplasia (MI) lesions during endoscopic examinations. In this example, the Raman spectroscopy system 200 included a near-infrared (NIR) diode laser (Aex = 785 nm) (maximum output: 300 mW, B&W TEK Inc.), a high-performance reflective imaging spectrograph. (Acton LS-785 f / 2, Princeton Instruments Inc.) equipped with an 830 gr / mm gold grid and a thermoelectrically cooled, NIR-optimized charge-coupled device (CCD) camera (PIXIS: 400BR-eXcelon, Princeton Instruments Inc.). System 200 obtained Raman tissue in vivo in the spectral range between 400 and 3600 cm-1 with a resolution of ~ 9 cm-1. A 1.9 m long 108 Raman fiber optic probe with an outer diameter of 1.8 mm for the delivery of laser light and the collection of Raman signals from epithelial tissue in vivo. The endoscopic Raman probe 108 designed for endoscopy includes 18 x 200 µm beveled collecting fibers (NA = 0.22) surrounding a central light delivery fiber (200 µm diameter, Na = 0.22). A 1.0 mm sapphire ball lens (NA = 1.78) is attached to the tip of the probe fiber to narrowly focus the excitation light onto the surface of gastric tissue, allowing effective collection of the Raman spectrum of the epithelial lining. The depth selective capability of the 200 Raman fiber optic spectroscopy system ensures shallow tissue interrogation (<200 pm) with microscopic probing volume (<0.02 mm2), thereby reducing interference and dilution of the signal from deeper voluminous tissues, while the epithelium associated with neoplastic onset and progression is selectively or preferentially interrogated. Atomic emission lines from mercury-argon spectral calibration lamps (HG-1 and AR-1, available from Ocean Optics, Inc., Dunedin, FL) were used for wavelength calibration. All calibrated wavelength spectra were corrected for system wavelength dependence using a tungsten calibration lamp (RS-10, EG&G, available from Gamma Scientific, San Diego, CA). The complete FP / HW Endoscopic Fiber Optic Raman System 200 can be controlled by a foot pedal and an intuitive software framework configured to provide feedback (eg, auditory and / or visual probabilistic feedback) to the gastroenterologist in real time.

The raw spectra were preprocessed using a third order Savitzky-Golay smoothing filter (a 3 pixel window width) to remove spectral noise. In the FP zone (between 800 and 1800 cm-1), a fifth order polynomial was found to be optimal to fit the AF background in the noise-smoothed spectrum, and then this polynomial was subtracted from the measured FP spectrum. to produce the Raman spectrum of FP tissue only. In the HW range (between 2800 and 3600 cm-1), a first-order polynomial fit was used to remove the AF background. The Raman spectra of ^f P / HW were normalized over the integrated area under the FP and HW ranges, to allow a better comparison of the spectral shapes and the relative intensities of the Raman band between normal gastric tissue and MI. All raw spectral data were processed online using software developed in the Matlab environment (Mathworks Inc., Natick, MA). Principal component analysis (PCA) and linear discriminant analysis (LDA) were implemented to develop robust diagnostic algorithms for the differentiation between normal and iM gastric tissues. Cross-validation, leaving a tissue site out, was used to evaluate the developed diagnostic models in an unbiased manner. Multiple Raman spectra (between 10 and 15) were obtained from each tissue site in 1 second, and the majority voting strategy was applied for the final classification. The diagnostic results could be displayed on a display device such as a computer screen in real time. Receiver operating characteristic (ROC) curves were also generated by successively changing the thresholds to determine correct and incorrect classifications for all tissues. All spectra preprocessing and multivariate statistical analysis were performed online using scripts written in the Matlab programming environment.

A total of 1412 Raman spectra were successfully obtained in vivo (i.e. normal (n = 1083 spectra) and MI (n = 329 spectra) from 115 sites (i.e. normal (n = 88 sites) and MI (n = 27 sites)) as confirmed by consensus histopathological examinations.

Figure 22A shows the mean standard deviation (SD) of in vivo Raman spectra ± 1 (shaded in light gray) measured (ie normal and MI). Prominent tissue Raman peaks with provisional assignments in the FP range are seen in:

• 875 cm-1, which is related to proteins v (C-C),

• 1004 cm-1, which is related to the aromatic ring v ^s (CC) of phenylalanine,

• 1078 cm-1, which is related to the v (C-C) of lipids,

• 1302 cm-1, which is related to the twisting and agitation of CH2 of lipids

• 1445 cm-1, which is related to strain 8 (CH2) of proteins and lipids, and

• 1655 cm-1, which is related to the amide I v (C = O) of proteins.

In addition, strong Raman peaks are also observed in the HW zone in

• 2885 cm-1, which is related to the symmetric and asymmetric stretching of lipid CH2s,

• 2940 cm-1, which is related to the stretching of CH3 in proteins,

• 3300 cm-1, which is related to amide A (NH stretching of proteins), and

• the broad Raman band of water (stretching vibrations of OHs reaching a maximum of ~ 3400 cm-1) which is related to the local conformation and interactions of OH bonds in the cellular and extracellular space of the tissue.

Figure 22B shows the difference spectra (i.e., MI - normal) ± 1 SD (shaded light gray) that resolve the distinctive spectral features (e.g., peak intensity, shift, and band broadening) associated with Mi transformation. , confirming the potential of FP / HW Raman spectroscopy for the early diagnosis of MI in endoscopy.

Figures 23A-B show representative hematoxylin and eosin sections (H and E) from corresponding tissue sites, using Raman endoscopy, including (a) normal gastric mucosa (x200 magnification); (b) extensive gastric intestinal metaplasia, in which the gastric epithelium contains apparent goblet cells (x100 magnification). Histopathological characterizations reveal the cellular and morphological characteristics of normal and intestinal metaplastic lesions in gastric tissue, while simultaneous FP / HW Raman endoscopy uncovers specific biochemical components (e.g., proteins, lipids, and water) of epithelial tissue at the level molecular.

To develop sophisticated multivariate diagnostic algorithms and compare the tissue diagnostic performance between the three different Raman techniques (i.e. FP, HW, and integrated FP / HW), PCA-LDA was implemented in conjunction with Student's t-test on spectra. Raman tissue in vivo obtained, to evaluate the elusive differences observed in the spectra of different types of tissues. The PCA-LDA diagnostic algorithms with cross-validation leaving a tissue site out were further developed based on significant diagnostic PCs (p <0.01), as shown in Figure 24, representing 45.6% (PC1), 33.6% (PC2), 4.2% (PC3), 3.1% (PC4) and 1.2% (PC5) of Raman spectral variations, respectively. The characteristics of different significant PCs are different; in particular, some PC characteristics, such as the peaks, valleys, and spectral shapes in Figure 24, are similar to those of tissue Raman spectral patterns. The first significant PC represents the largest variation within the spectral data sets (ie, 45.6%), while successive PCs describe characteristics that contribute to progressively smaller variations. The five PCs of diagnostic significance are then entered into the LDA model, along with a cross-validation technique leaving a tissue site out, for diagnosis and classification of gastric tissue.

Figure 25 shows results of the cross-validation classification (posterior probabilities) between normal and MI pathologies by modeling the PCA-LDA algorithm calculated for the Raman technique (a) FP, (b) HW and (c) FP / Integrated or simultaneous HW, respectively. Threshold lines (0.5) applied to the posterior probability scatter plots lead to diagnostic accuracies of 89.6% (103/115), 78.3% (90/115), and 91.3% (105/115) (sensitivities of 96.3% (26/27), 77.8% (21/27), 92.6% (25/27), and specificities of 87.5% (77 / 88), 78.4% (69/88), 90.9% (80/88)), respectively, for the integrated FP, HW and FP / HW Raman spectroscopic techniques. Results demonstrate that integrated FP / HW Raman spectroscopy works better for the diagnosis of gastric MI compared to FP-only or HW-only technique.

Figure 26 shows ROC curves generated for the integrated FP, HW, and FP / HW Raman techniques, revealing the relationships between diagnostic sensitivities and specificities of gastric MI identification. The areas of integration under the curves (AUC) are 0.94, 0.79, and 0.96, respectively, for the integrated FP, HW and FP / HW Raman techniques, again confirming that the FP / Integrated HW provides the best diagnostic performance for detection of gastric MI in vivo. Overall, the above results demonstrate the great potential of the simultaneous FP / HW Raman spectroscopic technique developed to improve the early diagnosis of gastric precancerous lesions in vivo during endoscopic examination.

Figure 27 shows a further embodiment of Raman spectra measured in the oral cavity in the FP only range. It should be noted that there is prominent interanatomic variability between different tissue sites in the oral cavity (eg, buccal and masticatory mucosa). In this embodiment, areas of large interanatomic variation, such as 956 cm-1 of hydroxyapatite and 1302 cm-1 and 1445 cm-1 of lipids, are excluded using the interval selection procedure. Therefore, variable / feature selection techniques can be used to reduce the effect of interanatomic variability in diagnostic algorithms by excluding spectral areas with large variation. This is an important point and will have applications in other organs, including the skin, the oral cavity, etc.

Figure 28 shows an example of how prediction can be applied prospectively in real time to new spectra after implementation of a probabilistic diagnostic model based on different spectral subzones.

This includes several steps, including fiber background subtraction calibration, truncation to spectral ranges, pre-processing, and finally, application of the predictor for disease classification.

This embodiment has several important advantages over the prior art: instead of using the entire broadband spectrum for diagnosis, this technique or procedure only uses a subset (for example, between 5 and 20%) of the supplementary information, greatly simplifying the diagnostic model. Second, the selection of variables provides qualitative insights into the biochemical and biomolecular basis of the disease. Third, the precision is significantly increased compared to spectral analysis based on the entire spectral range. Also, if certain spectral ranges have interferences or confounders (eg blood), etc., the effect of these can be efficiently reduced.

In one example, a patient scheduled for endoscopic screening for suspected esophageal reflux symptoms is subjected to the following:

• Conventional WLR / NBI / AFI endoscopic imaging is performed in the distal esophagus, showing an inconclusive appearance of large Barrett's segments suspected of precancer (ie, dysplasia).

• Subsequently, the broadband confocal Raman endoscope technique is applied to objectively direct the biopsies to the suspicious tissue segments.

• The confocal Raman classification is defined as “normal” (absence of pathology or gastritis), “low risk” (intestinal metaplasia) and “high risk” (dysplasia / cancer).

• The confocal Raman probe is placed against the tissue in the distal esophagus and the diagnosis is given in real time, based on complementary information extracted from the FP and HW area.

• The confocal Raman endoscope technique is focused on high-risk tissue sites that are then biopsied.

Embodiments according to the present invention demonstrate that simultaneous real-time high wave number (HW) and fingerprint (FP) confocal fiber optic Raman spectroscopy can be performed during in vivo patient screening. Fiberoptic confocal Raman spectroscopy discovers biomolecular and biochemical changes, such as proteins, DNA, lipids, and water found in the epithelium during colorectal carcinogenesis. The use of complementary biomolecular information from the FP and HW range improves the detection of anomalous growths compared to the use of the FP or HW only ranges. The FP and HW fiber optic confocal Raman spectroscopy technique has great potential to improve the detection and characterization of cancer and precancer. Using a subset of the entire spectrum has a remarkable advantage in processing and identifying abnormal growths in many different parts of the body. The FP and HW ranges have been especially identified to produce good results for the types of detection discussed herein. It will be appreciated that, for other types of detection, other ranges may be more useful.

Embodiments according to the present invention are not limited to biomedical Raman spectroscopy, but may also have application in other areas. These include, for example, fluorescence spectroscopy, elastic scattering spectroscopy, surface enhanced Raman spectroscopy, process analytical technology, water and environmental monitoring, pharmaceutical process control / drug supply, food industry, quality control industry, forensic medicine, etc. In such embodiments, the excitation energy provided by an illumination source is directed at a target sample (eg, a sample of a chemical, water or environmental material / substance, a pharmaceutical / drug or food sample, or other sample type), which need not include a fabric or be a fabric. Furthermore, such embodiments need not involve an endoscope or endoscopy. Reference to the term Raman herein is intended to include other types of spectroscopy, including those mentioned above.

Embodiments according to the present invention can serve as a diagnostic platform for any organs, such as the lung, upper and lower GI (eg, esophagus, stomach, colon and rectum), liver, bladder, head and neck (eg, nasopharynx, larynx, oral cavity), cervix, skin, bone, or any other place where a conventional endoscope, laparoscope, or arthroscope can be used.

Some of the Raman spectra of tissue obtained using certain embodiments in accordance with the present invention may suffer from endoscope illumination interference. This can make it difficult to diagnose abnormal growths. As a result, a further embodiment offers a solution to eliminate such interferences.

An endoscopic trimodal imaging system, according to an embodiment of the present invention, used to guide the confocal Raman fiber probe as described above, includes a 300 W short arc xenon light source, a gastrointestinal endoscope (GI) and a video signal processor. The xenon light source is coupled with different sets of filters to provide one or more different illumination lights for tandem trimodal endoscopic imaging (not shown). Filters can include, for example: WLR (red filter, between 585 and 655nm; green filter, between 500 and 575nm; blue filter, between 390 and 495nm), AFI (blue filter, between 390 and 470nm; green filter, between 540 and 560nm for reflective image normalization) and NBI (green filter, between 530 and 550nm; blue filter, between 390 and 445nm).

Reflected light or fluorescence emitted from the tissue is detected using two different CCDs mounted on the distal tip of the GI endoscope (not shown): one is a conventional CCD for WLR / NBI, and the other is a high sensitivity CCD for observation. by AFI. The endoscope's short arc xenon lamp emits continuous broadband light that covers UV / VIS / NIR rays with prominent discrete peaks in the NIR spectral range> 700 nm. Since the 785nm laser excitation Stokes Raman spectroscopy also falls into the NIR range, ambient xenon light can interfere with the tissue Raman signal and completely obscure in vivo tissue measurements and diagnostics. For this reason, endoscopic Raman diagnosis is conventionally performed with xenon illumination light turned off or dimmed, which is highly undesirable in clinical settings.

To eliminate interference from ambient xenon illuminations, an embodiment is proposed in which the integration of a hot mirror low-pass filter (cutoff at ~ 700nm, average transmission ~ 95% in visible range) in front of the source of Xenon light in the endoscope system, allows the elimination of environmental interference.

To demonstrate the application of this embodiment, Raman spectra are measured in vivo under different endoscopic lighting conditions. Figure 29A shows the Raman spectrum of non-contact tissue, obtained during endoscopy without the integrated hot mirror. Significant interferences from the xenon light can be discerned that completely bypass the Raman signal of the tissue. Figure 29B shows the Raman spectrum of tissue in non-contact mode after integration of a hot mirror low pass filter against the xenon lamp. As can be seen, xenon interferences are substantially eliminated. High resolution tissue Raman peaks can now be observed with tentative molecular assignments as follows:

• 853 cm-1, which is related to proteins v (C-C),

• 1004 cm-1, which is related to the aromatic ring vs (C-C) of phenylalanine,

• 1078 cm-1, which is related to the v (C-C) of lipids,

• 1265 cm-1, which is related to amide III v (C-N) and or (N-H) of proteins,

• 1302 cm-1, which is related to the twisting and agitation of CH3CH2 in proteins,

• 1445 cm-1, which is related to the deformation of the 5 (CH2) of proteins and lipids,

• 1655 cm-1, which is related to the amide I v (C = O) of proteins, and

• 1745 cm-1, which is related to the v (C = O) of lipids.

Raman spectra were measured with the probe in contact with the tissue. Figure 30A shows the Raman spectrum of tissue measured in contact mode in the absence of the hot mirror. Figure 30B shows the Raman spectrum of tissue in contact mode after integration of a hot mirror low pass filter against the xenon lamp. Overall, these results show that hot mirror integration can completely remove xenon environmental interferences from tissue Raman spectra in vivo in both non-contact mode and confocal Raman probe contact mode. Furthermore, it is evident from the figures that the xenon interference is more prominent in the non-contact mode of the Raman probe with the tissue, as more light is coupled to the probe head. Therefore, this embodiment is of particular value in organs such as the larynx / bronchi, where the non-contact mode is preferred due to stimulation of the cough reflex, or in patients with advanced tumors where contact of the Raman probe with tissue it can induce bleeding and risk of cancer cell spread.

It should be noted that filtering the illumination light as a guide using a hot mirror is not limited to endoscopic application, but is more general, to filter any light or imaging modality used to guide NIR tissue spectroscopy ( i.e. Surface Enhanced Raman Spectroscopy (SERS), NIR fluorescence or NIR reflectance) for internal or external organs (e.g. skin, cervix, bladder, stomach, esophagus, nasopharynx, larynx, oral cavity, colon, rectum, lung, etc.).

Therefore, the concept has been generalized in Figure 31, which shows a 2100 system. The 2100 system includes a 2102 probe, which can be brought into contact with a 2104 tissue. The system further includes a 2106 NIR laser , a spectrograph 2108 and a CCD 2110. An edge long pass filter 2112 is located between the probe and the spectrograph and a band pass filter 2114 is located between the probe and the laser. A 2116 PC and associated software are used to control the system. The tissue is also illuminated by a light source 2118 through a hot mirror low-pass filter 2120. It will be appreciated that this system can vary in terms of the elements that make up the system and the orientation and position of said elements, depending of the precise application for which this embodiment is used.

Instrument standardization is known to lag far behind the pace of advances in clinical Raman spectroscopy. Because tissue Raman spectroscopy relies on inherently weak, high-resolution peaks, the technology is very sensitive to instrumental changes. Therefore, it is imperative to develop and employ techniques to test / calibrate and standardize Raman endoscopy instrumentation prior to clinical measurements in human patients, to ensure that in vivo diagnosis is reliable and consistent between different Raman systems.

Figure 32 shows the basic principle of the Raman endoscope test before the application in patients, which comprises (i) start-up of the system, (ii) test of the Raman endoscopy system (iii) application of the Raman endoscope in patients. A new apparatus and technique or procedure is proposed to test the system and calibrate a fiber optic Raman endoscope. This includes an opto-mechanical device and a set of program instruction routines implemented in Raman software for automatic system testing and calibration in fiber optic Raman diagnostic applications. The calibration device consists of spectral calibration radiation sources (for example, mercury and argon lamps, NIR spectral emitters, laser light, etc.), a light intensity radiation source, a laser power meter, a fabric phantom, together with automated opto-mechanical parts (eg a stepper motor) and controllers for standardization of the Raman endoscopy system. The calibration device is an integral part of the clinical Raman endoscopy platform, but can also be used as a standalone device, for general testing and fiber optic calibration comprising a 2302 filter wheel driven by a stepper motor, such as the FW102C, Thorlabs Inc. Newton, NJ, USA, which has different 2304 samples embedded in it. The flexible fiber optic 2306 Raman probe can be mounted to the calibration fixture as shown schematically in Figure 33.

The rotation of the filter wheel has been synchronized with the excitation and acquisition of the laser in the clinical Raman software. Spectra are obtained from each sample on the filter wheel and stored. Steps performed in calibration / test routines comprise measurement of CCD characteristics (i.e. temperature, dark current), laser excitation power, and fused silica fiber background, a fluorescent material for calibration / testing response of the system, a material for wavelength calibration / test, and finally a tissue phantom. The measured signals can be used to recalibrate the parameters (eg, intensity response, wavelength accuracy, background noise, etc.) of the Raman endoscope system.

The routines and standard samples that have been integrated into the filter wheel of the calibration device will now be described, along with examples of the signals obtained. Figure 34 shows the procedure (2400) for testing the Raman endoscope prior to use on patients using the calibration device. Signal analysis (i.e. fault detection or calibration) can be performed after each measurement or after all measurements have been completed (as shown in this specific example). In a first case, a detector signal from the CCD is measured and the temperature is recorded (2402). This may include measuring multiple spectra (ie, at an exposure time between 0 and 1 seconds) of the signal intensity in the absence of laser excitation. Subsequently, the system verifies that the darkness current is less than a maximum value that was previously stored in a factory configuration file. This embodiment further includes ensuring that the coefficient of variation of the dark current, over a series of spectra, is less than a maximum value that was also previously stored in the configuration file.

The excitation power of the laser at the tip of the fiber optic Raman probe is also measured (2402) using an integrated laser power meter, to ensure that it is within a range that is less than the maximum allowable exposure limit of the skin of the American National Standards Institute (ANSI - American National Standards Institute) (which is set on a 785 nm laser beam). This embodiment includes ensuring that the laser power is less than a maximum value that was previously stored in the configuration file. The filter wheel contains an empty slot with a dark environment so that a background signal from the fused silica fiber probe can be measured (2404). Multiple spectra of the fiber probe bottoms are measured with laser excitation and different exposure times (eg, between 0.1 and 1.0 s). These spectra contain information about the status of the fiber optic probe. For example, if the fiber is damaged or the probe tip is contaminated, this could be reflected in an increase in Raman intensity or fluorescence.

An example of background signals for the broadband Raman endoscopy probe for two different fiber optic probes is shown in Figure 35. This embodiment further includes ensuring that the coefficient of variation, over a series of spectra, is less than a maximum value that was previously stored in the configuration file. The measured fiber optic probe background is also stored in memory, for further processing (i.e. subtraction) from the Raman spectra of tissue in vivo during real-time processing.

A fluorescent glass exhibiting a broad spectrum of stable fluorescence (e.g. chromium-doped borosilicate glass, green glass, kopp2412, etc.) is also measured and stored to test and / or calibrate response and collection efficiency. Raman endoscopy system (2406). Standard fluorescence glass is pre-calibrated at the factory, using a National Institute of Standards and Technology (NIST) tungsten lamp, according to the procedure detailed in Figure 36. System response can be compared with a response that was previously stored in the config file. This embodiment further includes ensuring that the coefficient of variation, over a series of spectra, is less than a maximum value that was previously stored in the configuration file. The spectrum of the response standard sample can be used to recalibrate the system to factory calibration. Figure 37 shows examples of calibration functions derived from a standard fluorescent glass (ie kopp2412) using 2 different fiber optic probes.

A material with well defined narrow Raman peaks is also measured (2408). An example is polystyrene, shown in Figure 38. The wavelength standard is used to assess whether there has been a deviation in the wavelength axis due to optical misalignment. The spectrum is compared (eg using correlation coefficients or peak identification, etc.) with a spectrum that was previously stored in the configuration file. This includes proof of resolution (i.e. full width of half the maximum (FWHM, Full Width of Half Maximum)) and verification of the peak positions. The spectrum of the material with narrow and well defined Raman peaks can also be used to realign the system to the previous factory alignment stored in the configuration file. Recalibration comprises polynomial mapping (eg, 3 to 5 order) between predefined peaks from the factory calibrated wavelength axis and corresponding peaks from the misaligned spectrum. An example is shown in Figure 39 showing the polynomial mapping of peaks after recalibration.

The filter wheel also contains a layered fabric phantom. The tissue phantom consists of a material with diffuse properties and / or layered tissue phantoms exhibiting well-known Raman peaks that are measured (2410). This embodiment further includes ensuring that the coefficient of variation, over a series of spectra, is less than a maximum value that was previously stored in the configuration file. In Figures 40A-B, an example is provided showing a comparison of Raman spectra obtained from the phantom sandwich of two layers of polystyrene and polyethylene using different Raman probes. The Raman spectra of the layered tissue phantoms are used to verify the depth selectivity of the fiber optic Raman probe as well as the quality of the spectrum. The ratio of the signal from the top to the bottom layer (eg, 1002 cm-1 and 1296 cm-1, respectively) is used as a qualitative and quantitative indicator of depth selectivity for a given fiber optic confocal probe. This embodiment includes ensuring that the depth selectivity is less than a maximum value that was previously stored in the configuration file.

After the various measurements have been made, analysis of the signals is carried out (2412). This is to identify problems such as fault detection, calibration problems, etc. If problems are detected, the system is defined as unsuccessful (2414). Alternatively, the system is defined as successful (2416) if no problems are found.

Another embodiment includes a set of program instructions or software framework integrated with, or as a GUI, that is capable of displaying and saving Raman data for each spectrum and lesion, in response to user selection (s) / input (s), when endoscopic Raman diagnosis is made. For example, if a first selection is made or a first button is pressed, the Raman data of a first lesion will be saved. If a second selection is made or a second button is pressed, the Raman data for a second lesion etc. will be saved. This system also includes the storage of patient information, integration time, laser excitation power, time, date, diagnosis, probe background signal, system calibration function, endoscopy video and a log file containing measurement details. An integrated foot pedal device may be provided for interconnection and storage of data by the clinician. If the footswitch is pressed, the data obtained will be stored as a first lesion. If pressed a second time, the data will be stored as a second injury, etc. The footswitch may also contain a safety pedal to turn the laser drive on and off.

Another embodiment includes a set of software framework or program instructions for displaying clinical Raman diagnosis in conjunction with recorded wide-field video imaging. A video endoscopic imaging system (ie, WLR / NBI / AFI) and a video recording system have been integrated into the Raman endoscope software. Both wide-field endoscopic video and in vivo Raman diagnosis of the imaged tissue can be displayed on a display device, such as a computer screen. The endoscopy video is time-synchronized with measured tissue Raman spectra. Therefore, spectral Raman diagnostics can be displayed simultaneously with video playback. This allows the clinician to track backward for each patient's precise correlation between the endoscopic Raman diagnosis and the suspicious tissue site sampled. This integrated Raman endoscopy and imaging software is a substantial enhancement to the clinical system, enabling the review of historical clinical data.

The limited dynamic range of CCDs and weak tissue Raman signals remain a challenge in endoscopic Raman applications, as in vivo tissues exhibit a variable degree of autofluorescence. For some tissues (eg stomach, lung, dorsal tongue, liver), detector saturation can occur in 0.1 seconds, or even less than 0.05 seconds. For these tissues, the Raman signal may have a prominent noise compared to other types of tissues (eg esophagus, nasopharynx, larynx, cervix, etc.). In general, the intensity of the tissue Raman signal scales linearly with the excitation power of the laser and the exposure time. The signal to noise ratio (S / N) is proportional to the square root of the exposure time. Currently, the user manually adjusts the exposure time or laser excitation powers for each spectrum obtained using the Raman endoscopy technique. This can be very impractical in real-time applications. Therefore, it is essential to define automated procedures to prevent CCD saturation and to ensure that the Raman spectrum is obtained with an optimal S / N ratio at times that are acceptable under clinical conditions. True real-time diagnosis is required for Raman endoscopy technology to gain wide acceptance in clinical medicine. The new procedure to automatically adjust laser excitation power, exposure time, and spectrum accumulations for uninterrupted real-time diagnostics during clinical measurements Raman with a high S / N ratio is invaluable in this search. Two automated techniques or procedures are set forth by way of representative example.

In a first technique or procedure, as shown with reference to Figure 41, the exposure time is fixed (for example, 0.1 seconds) (3102) and the laser driving power is assigned as variable and adjust as needed (3104). The Raman spectra are measured (3106) and the signal-to-noise ratio is determined (3108). The estimation of noise in the spectra is based on Fourier transform, differentiation or other procedures to quantify the noise level. If the Raman spectra have a bad signal-to-noise (S / N) ratio (3110), in the subsequent measurement, several Raman spectra are accumulated (eg, N = 2 ... 3) (3112) before a reading. If the CCD is saturated for a given spectrum (3114), the laser driving power of the subsequent Raman measurement is reduced by a number according to the degree of saturation (3116). If the spectrum is not saturated, it is used for diagnosis (3118). The degree of saturation is defined as a% of the saturated pixels. If the spectrum has been read with intensity counts (<70% or> 90% of the dynamic range) (3120) the driving power of the laser of the subsequent Raman measurement increases so that the signal intensity is within 70 at 90% dynamic range (3122). One embodiment imposes an upper limit on the driving power of the laser (eg, 25 mW). In another embodiment, the accumulation of various spectra (eg, n = 2) is done consistently. Background subtraction, preprocessing, outlier detection, disease prediction, and visualization can be performed (3124).

In a second technique or procedure as shown with respect to FIG. 42, the laser drive power is set (eg, 25 mW) (3202) and the exposure time is assigned as a variable (3204). The Raman spectra are measured (3206) and the signal-to-noise ratio is determined (3208). The noise estimation is based on Fourier transform, differentiation or any procedure to quantify the noise level. If the Raman spectrum shows a poor S / N ratio (3210), in the subsequent measurement, several Raman spectra are accumulated (eg, n = 2 ... 3) (3212) before reading. If the CCD is saturated (3214) for a given spectrum, the exposure time of the subsequent Raman measurement is reduced (3216) with a number according to the degree of saturation. If the spectrum is not saturated, it is used for diagnosis (3218). The degree of saturation is defined as the% of saturated pixels. If the spectrum has a low number of intensity counts (for example, <70% or> 90% of dynamic range) (3220), the exposure time of the subsequent Raman measurement is increased (3222) so that the intensity of the signal is within 70 to 90% of the dynamic range. One embodiment imposes an upper limit on the exposure time (eg 0.5 seconds). In another embodiment, the accumulation of various spectra is carried out consistently if the integration time is very low (eg, <0.1 s). Background subtraction, preprocessing, outlier detection, disease prediction, and visualization can be performed (3224).

The above two functions detailed with respect to Figures 36 and 37, respectively, are built into the operating software. Adjusting only the laser power has the advantage that the exposure time remains constant and can be very low (eg 0.1 seconds). The above procedures are of great value in clinical Raman endoscopy for real-time uninterrupted diagnosis with a high S / N ratio, but are also generalized to other modalities such as surface-enhanced Raman endoscopy (SERS), spectroscopy of reflectance and fluorescence. The general conception of automated adjustment of exposure time, laser excitation power, and accumulation to improve signal quality, prevent CCD saturation, and perform real-time diagnostics is an important feature of this embodiment of the present. invention. Another embodiment includes synchronizing the excitation of the laser with the acquisition. In other words, before obtaining the signal, the laser drive is turned on, and after obtaining, the laser drive is turned off.

Still another embodiment according to the present invention includes the ability to switch between different fiber optic probes and different organs. For each type of fiberoptic probe (for example, a confocal or volume probe), there is a database with specific diagnostic models for each organ (i.e., larynx, colon, nasopharynx, stomach, esophagus, oral cavity, skin, cervix, bladder, etc.). For example, if a user chooses the nasopharynx as the organ, one or more models belonging to the nasopharynx are loaded into the system. A representative database structure is shown in Figure 43. The model (s) may include multivariate diagnostic algorithms and / or multivariate outlier detection models.

It will be appreciated that the various embodiments described above are not intended to be limiting in their interpretation. Rather, the invention is defined in the appended claims. Some of the processes can be implemented in hardware, software or any combination of the same.

List of reference signs

100 system - fiber optic confocal Raman spectroscope

102 diode laser

104 imaging spectrograph

105 elongated cylindrical body

106 camera

108 probe

110 fiber optic

112 band pass filter

114 long pass filter

116 tip

124 microcontroller system

200 appliance

202 diode laser

204 reflective spectrograph

206 camera

208 fiber optic confocal Raman probe

Deoxyribonucleic acid DNA

HW high wave number regime

FP footprint

NIR Near Infrared

GUI Graphical user interface

PCA Principal Component Analysis

SVM support vector machine

LR logistic regression

LDA linear discriminant analysis

Partial least squares PLS

DA discriminant analysis

LV latent variables

PEG polyethylene glycol

ACO ant colony optimization

CART Classification Trees and AdaBoost Adaptive Boost Regression

H and E hematoxylin and eosin

ANOVA analysis of variance

LSD less significant differences

ROC receiver operating characteristics AUC area under curves

SD standard deviation

MI intestinal metaplasia

CCD charge coupled device

Gastrointestinal GI

SERS surface enhanced Raman spectroscopy NIST National Institute of Standards and Technology FWHM full width of half maximum

Claims (11)

1. Raman spectroscopy apparatus (2100), comprising: a first illumination source (2106), configured to direct a collimated illumination output to a tissue (2104); a Raman spectrograph (2108), configured to simultaneously detect fingerprint (FP) and high wave number (HW) Raman spectra from tissue scattered illumination; and a computerized control and analysis module (2116), comprising at least one processing unit and memory that stores program instructions executable by the at least one processing unit, to analyze discrete spectral sub-ranges of the Raman spectra in FP and HW wave number ranges to identify a match to specific peaks in the detected Raman spectra of one or both wave number ranges, characterized in that it also comprises: an additional illumination source (2118) configured to emit additional illumination to the tissue while the first illumination source directs the collimated illumination to the tissue and a hot mirror filter (2120) configured to compensate for illumination interference between the collimated illumination emitted by the first illumination source (2106) and the additional illumination emitted by the additional illumination source (2118). 2. Apparatus (2100), according to claim 1, characterized in that the Raman spectrograph (2108) has a single broadband diffraction grating, and in which the apparatus (2100) further comprises a probe (2102), to transmit the collimated illumination to the tissue and return the Raman spectra of the tissue to the spectrograph (2108). 3. Apparatus (2100) according to claim 2, characterized in that the computerized control and analysis module (2116) includes program instructions executable by at least one processing unit to diagnose abnormal growth based on the coincidence . Apparatus (2100) according to claim 2, characterized in that the probe (2102) comprises a confocal fiber optic probe and, optionally, an endoscope that has an elongated cylindrical body that has an instrument channel inside which it is housed. the probe (2102). Apparatus (2100) according to claim 2, characterized in that the computerized control and analysis module (2116) includes program instructions executable by the at least one processing unit to dynamically adjust the power of the collimated illumination. Apparatus (2100) according to claim 2, characterized in that the computerized control and analysis module (2116) includes program instructions executable by the at least one processing unit to dynamically adjust the time during which the tissue is exposed. to collimated illumination. Apparatus (2100) according to claim 2, characterized in that it further comprises a calibration apparatus configured to standardize the probe (2102) or the entire Raman apparatus with respect to at least one calibration reference. 8. Raman spectroscopy procedure, performed by a Raman spectroscopy apparatus (2100), comprising the procedure: directing collimated illumination emitted by a first illumination source (2106) to a tissue (2104); simultaneously detecting by means of a Raman spectrograph (2108) of fingerprints (FP) and high wave number (HW), Raman spectra of illumination scattered by the tissue; analyze discrete spectral sub-ranges in the Raman spectra detected in the FP and HW wavelength ranges; and identifying a match with one or more reference markers in one or both wave number ranges; characterized in that it also comprises: directing additional illumination to the tissue using an additional illumination source (2118), while directing the collimated illumination emitted by the first illumination source (2106) to the tissue; and compensating for the illumination interference between the collimated illumination emitted by the first illumination source (2106) and the additional illumination emitted by the additional illumination source (2118) using a hot mirror filter (2120). 9. Method according to claim 8, characterized in that the simultaneous detection of FP and HW Raman spectra comprises diffraction illumination in both FP and HW wavelength ranges using a single broadband diffraction grating. 10. The method according to claim 8, characterized in that it further comprises dynamically adjusting the power of the lighting and / or adjusting the time during which the fabric is exposed to lighting. 11. The method according to claim 8, characterized in that it further comprises performing a calibration or standardization procedure to standardize the Raman apparatus with respect to at least one calibration reference, before illuminating the tissue.